J. Am. Chem. SOC.1995,117, 2947-2948
Sequential One-Pot Aldol Reactions Catalyzed by 2-Deoxyribose-5-phosphateAldolase and Fructose-1,6-diphosphateAldolase Hanie J. M. Gijsen and Chi-Huey Wong* Department of Chemistry, The Scripps Research Institute 10666 North Torrey Pines Road, La Jolla, Califomia 92037 Received December 8, 1994
Several aldolases have been shown to be useful catalysts for asymmetric aldol condensation. The enzymes 2-deoxyribose5-phosphate aldolase (DERA, EC 4.1 .2.4)2-5 and fructose-1,6diphosphate aldolase from rabbit muscle (RAMA, EC 4.1.2.13)' have been reported to accept a wide variety of acceptor substrates and were used in the synthesis of a variety of carbohydrate derivatives. In this paper we describe the coupling of the DERA- and RAMA-catalyzed sequential aldol reactions to form a variety of 5-deoxy 7-carbon carbohydrate derivatives. Also it was observed for the first time that the reactions in this enzymatic process were not 100% stereoselective at the reaction center for both enzymes. The DERA-catalyzed aldol condensation between a substrate aldehyde 1 and acetaldehyde produces a new C2-unsubstituted aldehyde 2 that itself can be a substrate for a subsequent aldol reaction. Especially if this newly formed aldehyde cannot form a stable hemiacetal, a second condensation with acetaldehyde can take place as reported recently! RAMA uses dihydroxyacetone phosphate 3 (DHAP) as a donor substrate and has been shown to accept several C2-unsubstituted aldehydes as acceptor substrates.'S6 When the two aldolases and the right substrates are combined, the aldehyde formed after the DERA-catalyzed aldol condensation might undergo a second aldol condensation with DHAP catalyzed by RAMA to give a 5-deoxy ketose 4. Similar to the sequential aldol reaction catalyzed by DERA, the driving force behind this reaction will be the formation of a stable hemiketal product. In Scheme 1 the different reactions that can take place in the reaction mixture are depicted. To prevent the formation of products of type 5 from only M A - c a t a l y z e d reactions it will be necessary to use substrates that cannot form stable hemiketals after aldol reaction with DHAP. For initial experiments the substituted acetaldehydes were chosen with R = MeO, N3, MOMO, and C1. Hydroxyacetaldehyde (glycolaldehyde) was chosen for the study of a substrate that is capable of forming a stable hemiacetal or ketal after only a DERA- or RAMA-catalyzed reaction, respectively. If the reaction is continued long enough, an equilibrium might be established with the expected 5-deoxy ketose 4 as the major product. The ratio of the original substrates is of course an important factor. Too much acetaldehyde gives side products 6 derived from a sequential DERA-catalyzed but an excess of both initial acceptor aldehyde and acetaldehyde is necessary to get a high concentration of second acceptor aldehyde 2 for the RAMA-catalyzed reaction. It was found that
'
(1) Wong, C.-H.; Whitesides, G. M. Enzymes in Synfhefic Organic Chemistly; Pergamon: Oxford, 1994; Chapter 4. (2) Barbas, C. F., III;Wang, Y.-F.;Wong, C.-H.J . Am. Chem. SOC.1990, 112, 2013. (3) Chen, L.; Dumas, D. P.; Wong, C.-H. J. Am. Chem. SOC. 1992, 114, 741. (4) Gijsen, H. J. M.; Wong, C.-H. J . Am. Chem. SOC. 1994, 116, 8422. ( 5 ) Wong, C.-H.; Garcia-Junceda, E.; Chen, L.; Blanco, 0.; Gijsen, H. J. M.; Steensma, D. H. Submitted. (6) (a) Durrwachter, J. R.; Wong, C.-H. J . Org. Chem. 1988, 53,4175. (b) Bednarski, M. D.; Simon, E. S.; Bischofberger, N.; Fessner, W.-D.; Kim, M.-J.; Lees, W.; Saito, T.; Waldmann, H.; Whitesides, G. M. J . Am. Chem. SOC. 1989, 111, 627. (c) Liu, K. K.-C.; Pederson, R. L.; Wong, C.-H. J. Chem. Soc., Perkin Trans. I 1991, 2669. (d) Fessner, W.-D.; Walter, C. Angew. Chem., Int. Ed. Engl. 1992, 31, 614. (7) The Occurrence of a sequential DERA-catalyzed reaction can be diminished by using less DERA.
2947
a ratio of initial acceptor aldehyde:acetaldehyde:DHAPof 2:3: 1 gave reasonable yields of 5-deoxy ketoses 4. Since all the reactions in Scheme 1 took place, other minor products of these reactions were present in the reaction mixture to some extent. Additionally, NMR analysis of these crude mixtures revealed the presence of small quantities of two other 5-deoxy ketoses which were apparently diastereomers of the main products 4. These diastereomers were easily separated from the main product by flash chromatography (SiOz). Further removal of other impurities with Dowex 50(H+) in methanol, followed by flash chromatography (SiOa), provided the pure methyl glycosides 7. Scheme 2 depicts the different products. Yields of products were determined on the basis of their isolated methyl glycosides. Although diastereomers 9 and 10 were observed for all the substrates tested, they could be separated from each other only for methoxyacetaldehyde (la) and azidoacetaldehyde (lb). The configuration of products 9 and 10 was determined via extensive NMR homodecoupling experiments. The substrate glycolaldehyde (le) gave a low yield of product 8e, as expected because of the formation of product 5 (R = OH), which can form a stable hemiketal. A reasonable yield of 8e could be obtained by incubating acetaldehyde and glycolaldehyde with DERA for 2 days, before adding RAMA and DHAP. Initial buildup of 2 (R = OH)4made the subsequent formation of 8e in 13% yield possible. The main product formed in the reaction with the substrate chloroacetaldehyde (Id) was different from the expected 8d. Apparently dehydrohalogenation of the initially formed product takes place, giving compound 11 as the final product in 26% yield (Scheme 3). The same anhydro compound was formed from 8c and 8e after treatment with MeOWDowex 50(H+). All reactions catalyzed by DERA and RAMA reported previously have been observed or assumed to take place with 100% stereoselectivity, giving products with 3 s stereochemistry in the case of DERA and 3S,4R stereochemistry for the RAMAcatalyzed reaction.' This stereochemistry will give products 8 with three hydroxyl groups in axial positions. Side product 9 has the inverse stereochemistry at C4, one of the stereocenters formed in the RAMA-catalyzed reaction. Side products 10 has the inverse stereochemistry at C6, formed in the DERAcatalyzed reaction. Both side products 9 and 10 have less 1,3 diaxial interaction than 8 and are thus thermodynamically more stable products. Apparently, since all reactions are reversible and are run for a long time (thermodynamically controlled conditions), some inversion in the normal stereochemistry of both DERA and RAMA takes place. Inversion of the stereochemistry at C4 by other DHAP dependent aldolases has been observed,8and similar, but more pronounced, inversion has also been reported for N-acetylneuraminic acid a l d ~ l a s e .For ~ DERA and RAMA, this is the first occasion on which the catalyzed reactions have been observed not to proceed fully stereospecifically at the reaction center. Several experiments were carried out to prove that the inversion of stereochemistry at C4 and C6 was caused by DERA and RAMA, respectively. Incubation of substrates of type 1 or 2 with acetaldehyde and DHAP, without adding aldolases, did not give any detectable products of type 8, 9, or 10. The 3S,4S stereochemistry in product 9 is the usual stereochemistry (8) (a) Fessner, W.-D.; Sinerius, G.; Schneider, A.; Dreyer, M.; Schulz, G. E.; Badia, J.; Aguilar, J. Angew. Chem., In?. Ed. Engl. 1991, 30, 555. (b) Fessner, W.-D.;Eyrisch, 0.Angew. Chem., Int. Ed. Engl. 1992,31,56. (c) Eyrisch, 0.;Sinerius, G.; Fessner, W.-D. Carbohydr. Res. 1993, 238, 287. (9) (a) AugC, C.; Gautheron, C.; David, S. Tefrahedron 1990, 46, 201. (b) Gautheron, C.; Ichikawa, Y.; Wong, C.-H. J . Am. Chem. SOC.1991, 113,7816. (c) Lin, C.-H.;Sugai, T.;Halcomb, R. L.; Ichikawa, Y.; Wong, C.-H. J . Am. Chem. SOC.1992, 114, 10138. (d) Kragl, U.; Godde, A,; Wandrey, C.; Lubin, N.; AugC, C. J . Chem. Soc., Perkin Trans. I 1994, 119.
0002-786319511517-2947$09.00/00 1995 American Chemical Society
Communications to the Editor
2948 J. Am. Chem. SOC.,Vol. 117, No. IO, 1995 Scheme 1 OH 0
DERA
0 OPO? OH
2
1
OH 0 R+OPO,Z.
,
T
OH O
OH
HO
4
3
H
OH 5
6
Scheme 2" 0
0
0
DERA, RAMA
R OH
Pase
OH
OH
OH
8a: 45%
9a: 9% 9b: 3% 9c: c 2% 9d: c 3% 9e: 1%
HO
la: R = OMe lb:R=N3 IC:R = OMOM Id: R = C1 le: R = OH
HO R*
8b: 34% 8c: 23% 8d: < 5% 8e: 13%
H R *
OH HO
loa: 3% lob: 2% 1oc: < 1% 1Od: < 2% 10e: < 1%
a Reactions were run for 6 days at pH 7.3 with 3 mmol of 1, 4.5 mmol of acetaldehyde, 1.5 mmol of DHAP, 670 units of DERA, and 320 units (added in 3 portions) of RAMA in 30 mL total volume. Yields of products are based on isolated methyl glycosides or 11.
Scheme 3"
a
(a) Enzymatic Incubation; (b) Dowex 50(Hf), MeOH.
Scheme 4" 0
OH 0
12
OBz OBz
aOH: 96% l3 $OH: 4%
(a) Acetaldehyde, DERA; (b) NaB&; (c) BzOC1, pyridine.
formed by tagatose-I ,6-diphosphate aldolase (TDP-aldolase). To exclude the possibility that 9 was formed by other contaminating enzymes like TDP-aldolase,'O the two aldol reactions were carried out separately. Product 12 was isolated after DERA-catalyzed aldol condensation of Id with acetaldehyde (Scheme 4).4 After incubation of 12 with DHAP and RAMA,'O a mixture of diastereomers similar to that from the one-pot reaction was obtained. Incubation of 12 with DHAP and crude DERA did not give any detectable products of type 8, 9, or 10. The stereoselectivity of the DERA-catalyzed reaction was tested further by reducing product 12, followed by benzoylation of the obtained diol to give 13 (Scheme 4). The same product was synthesized as a racemic mixture by reaction of 3-butenol with NaOCl followed by benzoylation. Separation of the two enantiomers on HPLC, using a Daicel OD-H column, showed the presence of 96% of the 3 s enantiomer and 4% of the 3R enantiomer, corresponding to an ee of 92%, indicating that at least for some substrates the DERA-catalyzed aldol condensation (10) Contaminationof RAMA with TDP-aldolasecan be excluded since, except for FDP-aldolase, no other DHAP dependent aldolases are present in mammalian tissues.
is not 100% selective. RAMA-catalyzed condensationsbetween substrates of type 2 (Scheme 1) with the 3R stereochemistry are thermodynamically much more favorable as has been shown by the selective formation of products of type 10, when racemic 2 is used as substrate.6 Finally, chemical inversion of the stereochemistry at C4 and C6 via retro-aldoUaldo1 reactions was excluded as follows. Product 12 and the crude phosphate-product mixture obtained from 2a were kept under reaction conditions in the absence of enzymes for 1 week. After isolation and derivatization, the ee of 12 and the ratio of compounds 8a, 9a and 10a had not changed. In summary, this report describes a new multienzyme system for sequential aldol addition reactions. Due to the reversible formation of thermodynamically highly disfavored products of type 8 and the long reaction times, some inversion of the usual stereochemistry of both DERA and RAMA takes place. Since the side products are only a minor fraction of the major product, the present method of coupling the DERA- and RAMAcatalyzed aldol condensation provides a good, one-pot method for making 5-deoxy ketose derivatives 8, using simple starting materials. Further studies are in progress to improve the yield of these reactions and to couple DERA with other aldolases. Acknowledgment. We thank the NIH (GM 44154) and Quest Intemational for financial support of this research. Supplementary Material Available: Detailed Procedure for the synthesis of the methyl glycosides of 8, 9 and 10. 'H and 13C NMR spectra for compounds 8a,b, 9a,b, and 10a,b (methyl glycosides) and 812, 8e, 11, and 13 (4 pages). This material is contained in many libraries on microfiche, immediately follows this article in the microfilm version of the joumal, can be ordered from the ACS, and can be downloaded from the Intemet; see any current masthead page for ordering information and Intemet access instructions. JA9439635
